A new study published in the Journal of Hydrology presents a machine learning model developed by researchers at Concordia University to improve flood evacuation protocols.
Muhammad Almetwalli Ahmed, a doctoral candidate, and Samuel Lee, professor and chair of the university's Department of Civil Engineering, are using artificial intelligence to more accurately predict short-term river flows, critical data for evacuations.
Read more:
Atmospheric river: what is it and the largest in the world? Baby-shaped storm recorded in Africa by satellite Climate crisis: new technology uses AI to predict storms Study river flow in detail and predict more effective floods - Image: Hydrology How a Model-Driven Learning Model Works. historical data from hydrometric stations and new climate parameters such as precipitation, temperature and humidity. The researchers focused on measuring convection, which measures the speed of water movement between two stations on the Ottawa River. The research used decades of data collected by the Government of Canada and tested the model with data from other regions of the United States, such as the Pond and Missouri rivers. The model provides an accurate estimate of daily runoff and, in particular, helps predict real-time flow, up to 24 hours of water flow required for effective evacuation. The method, which uses nine forecasters (seven climatological and two historical), is adjusted according to the forecast period. Over time, this model will be operational and open to the public, predicting water levels in real time, similar to weather forecasts.
Ahmad's idea is for the government to use the model as a tool to plan evacuations, optimize transport logistics and save lives and property during floods.
The research could help authorities evacuate civilians more quickly - Image: humpheria/Shutterstock